xref: /linux/scripts/spdxcheck.py (revision dfab99544c55378c1692c585e5e4df88dc7944c3) 
1#!/usr/bin/env python
2# SPDX-License-Identifier: GPL-2.0
3# Copyright Thomas Gleixner <tglx@linutronix.de>
4
5from argparse import ArgumentParser
6from ply import lex, yacc
7import locale
8import traceback
9import sys
10import git
11import re
12import os
13
14class ParserException(Exception):
15    def __init__(self, tok, txt):
16        self.tok = tok
17        self.txt = txt
18
19class SPDXException(Exception):
20    def __init__(self, el, txt):
21        self.el = el
22        self.txt = txt
23
24class SPDXdata(object):
25    def __init__(self):
26        self.license_files = 0
27        self.exception_files = 0
28        self.licenses = [ ]
29        self.exceptions = { }
30
31# Read the spdx data from the LICENSES directory
32def read_spdxdata(repo):
33
34    # The subdirectories of LICENSES in the kernel source
35    license_dirs = [ "preferred", "other", "exceptions" ]
36    lictree = repo.head.commit.tree['LICENSES']
37
38    spdx = SPDXdata()
39
40    for d in license_dirs:
41        for el in lictree[d].traverse():
42            if not os.path.isfile(el.path):
43                continue
44
45            exception = None
46            for l in open(el.path).readlines():
47                if l.startswith('Valid-License-Identifier:'):
48                    lid = l.split(':')[1].strip().upper()
49                    if lid in spdx.licenses:
50                        raise SPDXException(el, 'Duplicate License Identifier: %s' %lid)
51                    else:
52                        spdx.licenses.append(lid)
53
54                elif l.startswith('SPDX-Exception-Identifier:'):
55                    exception = l.split(':')[1].strip().upper()
56                    spdx.exceptions[exception] = []
57
58                elif l.startswith('SPDX-Licenses:'):
59                    for lic in l.split(':')[1].upper().strip().replace(' ', '').replace('\t', '').split(','):
60                        if not lic in spdx.licenses:
61                            raise SPDXException(None, 'Exception %s missing license %s' %(ex, lic))
62                        spdx.exceptions[exception].append(lic)
63
64                elif l.startswith("License-Text:"):
65                    if exception:
66                        if not len(spdx.exceptions[exception]):
67                            raise SPDXException(el, 'Exception %s is missing SPDX-Licenses' %excid)
68                        spdx.exception_files += 1
69                    else:
70                        spdx.license_files += 1
71                    break
72    return spdx
73
74class id_parser(object):
75
76    reserved = [ 'AND', 'OR', 'WITH' ]
77    tokens = [ 'LPAR', 'RPAR', 'ID', 'EXC' ] + reserved
78
79    precedence = ( ('nonassoc', 'AND', 'OR'), )
80
81    t_ignore = ' \t'
82
83    def __init__(self, spdx):
84        self.spdx = spdx
85        self.lasttok = None
86        self.lastid = None
87        self.lexer = lex.lex(module = self, reflags = re.UNICODE)
88        # Initialize the parser. No debug file and no parser rules stored on disk
89        # The rules are small enough to be generated on the fly
90        self.parser = yacc.yacc(module = self, write_tables = False, debug = False)
91        self.lines_checked = 0
92        self.checked = 0
93        self.spdx_valid = 0
94        self.spdx_errors = 0
95        self.curline = 0
96        self.deepest = 0
97
98    # Validate License and Exception IDs
99    def validate(self, tok):
100        id = tok.value.upper()
101        if tok.type == 'ID':
102            if not id in self.spdx.licenses:
103                raise ParserException(tok, 'Invalid License ID')
104            self.lastid = id
105        elif tok.type == 'EXC':
106            if id not in self.spdx.exceptions:
107                raise ParserException(tok, 'Invalid Exception ID')
108            if self.lastid not in self.spdx.exceptions[id]:
109                raise ParserException(tok, 'Exception not valid for license %s' %self.lastid)
110            self.lastid = None
111        elif tok.type != 'WITH':
112            self.lastid = None
113
114    # Lexer functions
115    def t_RPAR(self, tok):
116        r'\)'
117        self.lasttok = tok.type
118        return tok
119
120    def t_LPAR(self, tok):
121        r'\('
122        self.lasttok = tok.type
123        return tok
124
125    def t_ID(self, tok):
126        r'[A-Za-z.0-9\-+]+'
127
128        if self.lasttok == 'EXC':
129            print(tok)
130            raise ParserException(tok, 'Missing parentheses')
131
132        tok.value = tok.value.strip()
133        val = tok.value.upper()
134
135        if val in self.reserved:
136            tok.type = val
137        elif self.lasttok == 'WITH':
138            tok.type = 'EXC'
139
140        self.lasttok = tok.type
141        self.validate(tok)
142        return tok
143
144    def t_error(self, tok):
145        raise ParserException(tok, 'Invalid token')
146
147    def p_expr(self, p):
148        '''expr : ID
149                | ID WITH EXC
150                | expr AND expr
151                | expr OR expr
152                | LPAR expr RPAR'''
153        pass
154
155    def p_error(self, p):
156        if not p:
157            raise ParserException(None, 'Unfinished license expression')
158        else:
159            raise ParserException(p, 'Syntax error')
160
161    def parse(self, expr):
162        self.lasttok = None
163        self.lastid = None
164        self.parser.parse(expr, lexer = self.lexer)
165
166    def parse_lines(self, fd, maxlines, fname):
167        self.checked += 1
168        self.curline = 0
169        try:
170            for line in fd:
171                line = line.decode(locale.getpreferredencoding(False), errors='ignore')
172                self.curline += 1
173                if self.curline > maxlines:
174                    break
175                self.lines_checked += 1
176                if line.find("SPDX-License-Identifier:") < 0:
177                    continue
178                expr = line.split(':')[1].strip()
179                # Remove trailing comment closure
180                if line.strip().endswith('*/'):
181                    expr = expr.rstrip('*/').strip()
182                # Special case for SH magic boot code files
183                if line.startswith('LIST \"'):
184                    expr = expr.rstrip('\"').strip()
185                self.parse(expr)
186                self.spdx_valid += 1
187                #
188                # Should we check for more SPDX ids in the same file and
189                # complain if there are any?
190                #
191                break
192
193        except ParserException as pe:
194            if pe.tok:
195                col = line.find(expr) + pe.tok.lexpos
196                tok = pe.tok.value
197                sys.stdout.write('%s: %d:%d %s: %s\n' %(fname, self.curline, col, pe.txt, tok))
198            else:
199                sys.stdout.write('%s: %d:0 %s\n' %(fname, self.curline, col, pe.txt))
200            self.spdx_errors += 1
201
202def scan_git_tree(tree):
203    for el in tree.traverse():
204        # Exclude stuff which would make pointless noise
205        # FIXME: Put this somewhere more sensible
206        if el.path.startswith("LICENSES"):
207            continue
208        if el.path.find("license-rules.rst") >= 0:
209            continue
210        if not os.path.isfile(el.path):
211            continue
212        with open(el.path, 'rb') as fd:
213            parser.parse_lines(fd, args.maxlines, el.path)
214
215def scan_git_subtree(tree, path):
216    for p in path.strip('/').split('/'):
217        tree = tree[p]
218    scan_git_tree(tree)
219
220if __name__ == '__main__':
221
222    ap = ArgumentParser(description='SPDX expression checker')
223    ap.add_argument('path', nargs='*', help='Check path or file. If not given full git tree scan. For stdin use "-"')
224    ap.add_argument('-m', '--maxlines', type=int, default=15,
225                    help='Maximum number of lines to scan in a file. Default 15')
226    ap.add_argument('-v', '--verbose', action='store_true', help='Verbose statistics output')
227    args = ap.parse_args()
228
229    # Sanity check path arguments
230    if '-' in args.path and len(args.path) > 1:
231        sys.stderr.write('stdin input "-" must be the only path argument\n')
232        sys.exit(1)
233
234    try:
235        # Use git to get the valid license expressions
236        repo = git.Repo(os.getcwd())
237        assert not repo.bare
238
239        # Initialize SPDX data
240        spdx = read_spdxdata(repo)
241
242        # Initilize the parser
243        parser = id_parser(spdx)
244
245    except SPDXException as se:
246        if se.el:
247            sys.stderr.write('%s: %s\n' %(se.el.path, se.txt))
248        else:
249            sys.stderr.write('%s\n' %se.txt)
250        sys.exit(1)
251
252    except Exception as ex:
253        sys.stderr.write('FAIL: %s\n' %ex)
254        sys.stderr.write('%s\n' %traceback.format_exc())
255        sys.exit(1)
256
257    try:
258        if len(args.path) and args.path[0] == '-':
259            stdin = os.fdopen(sys.stdin.fileno(), 'rb')
260            parser.parse_lines(stdin, args.maxlines, '-')
261        else:
262            if args.path:
263                for p in args.path:
264                    if os.path.isfile(p):
265                        parser.parse_lines(open(p, 'rb'), args.maxlines, p)
266                    elif os.path.isdir(p):
267                        scan_git_subtree(repo.head.reference.commit.tree, p)
268                    else:
269                        sys.stderr.write('path %s does not exist\n' %p)
270                        sys.exit(1)
271            else:
272                # Full git tree scan
273                scan_git_tree(repo.head.commit.tree)
274
275            if args.verbose:
276                sys.stderr.write('\n')
277                sys.stderr.write('License files:     %12d\n' %spdx.license_files)
278                sys.stderr.write('Exception files:   %12d\n' %spdx.exception_files)
279                sys.stderr.write('License IDs        %12d\n' %len(spdx.licenses))
280                sys.stderr.write('Exception IDs      %12d\n' %len(spdx.exceptions))
281                sys.stderr.write('\n')
282                sys.stderr.write('Files checked:     %12d\n' %parser.checked)
283                sys.stderr.write('Lines checked:     %12d\n' %parser.lines_checked)
284                sys.stderr.write('Files with SPDX:   %12d\n' %parser.spdx_valid)
285                sys.stderr.write('Files with errors: %12d\n' %parser.spdx_errors)
286
287            sys.exit(0)
288
289    except Exception as ex:
290        sys.stderr.write('FAIL: %s\n' %ex)
291        sys.stderr.write('%s\n' %traceback.format_exc())
292        sys.exit(1)
293